martin [Sat, 30 Aug 2008 20:45:31 +0000 (20:45 +0000)]
replace TRUNC_PAGE by trunc_page and ROUND_PAGE by round_page
plus cast to vaddr_t where necassary
from Miod's todo list
ok miod@
miod [Sat, 30 Aug 2008 20:13:03 +0000 (20:13 +0000)]
Code refactorization and state structure changes in preparation of asynchronous
abort commands, needed to eventually support command timeouts for non-polled
commands... no functional change.
martin [Sat, 30 Aug 2008 18:58:24 +0000 (18:58 +0000)]
no need to include uvm/uvm.h twice
fgsch [Sat, 30 Aug 2008 10:41:38 +0000 (10:41 +0000)]
use a bitmap to store the features. simplify the code a bit and allow for
future work. ok av@ a similar diff.
brad [Sat, 30 Aug 2008 08:16:13 +0000 (08:16 +0000)]
Move the link state check back up to the top of the various status
functions right after reading the BMSR as it was before the addition
of the fiber support and simplify brgphy_fiber_status() a bit.
ok dlg@
brad [Sat, 30 Aug 2008 07:41:51 +0000 (07:41 +0000)]
garbage collect the no longer used GEM_IS_APPLE macro.
brad [Sat, 30 Aug 2008 07:39:12 +0000 (07:39 +0000)]
Fix an issue initially reported by drahn@ with gem(4) where these
drivers can report 2-3x times the number of actual packets being
transmitted by only incrementing the counter for descriptors with
buffers being freed.
ok kettenis@
brad [Sat, 30 Aug 2008 07:25:11 +0000 (07:25 +0000)]
regen
brad [Sat, 30 Aug 2008 07:24:32 +0000 (07:24 +0000)]
Add PCI ids for Intel E1000 82576 based boards.
From FreeBSD
brad [Sat, 30 Aug 2008 04:51:45 +0000 (04:51 +0000)]
regen
brad [Sat, 30 Aug 2008 04:50:50 +0000 (04:50 +0000)]
Add the vendor ID for Techsan Electronics and PCI ids for their
Sky2PC DVB cards.
From jbg@
jsg [Sat, 30 Aug 2008 02:46:56 +0000 (02:46 +0000)]
Remove the header that exposed the old wdt ioctl interface to userland,
wdt was converted to use the common watchdog code some time ago so this
is no longer needed.
"yes, we should remove it" mbalmer@
brad [Sat, 30 Aug 2008 01:16:30 +0000 (01:16 +0000)]
- Remove passing a redundant --no-split flag on the command line
for makeinfo.
- Remove some white space.
ok beck@ biorn@
brad [Sat, 30 Aug 2008 01:10:13 +0000 (01:10 +0000)]
Mention Family 11h in the comment.
brad [Sat, 30 Aug 2008 01:09:08 +0000 (01:09 +0000)]
Correct the last commit. There is no such thing as the K11
microarchitecture (yet). Family 11h (Griffin) is a K8
microarchitectre CPU with some bits from K10 such as power
management and sensor interface.
brad [Fri, 29 Aug 2008 23:28:34 +0000 (23:28 +0000)]
Disable the use of Jumbo frames on the first generation chips (82542).
The exact details are unknown (possibly hw errata?) but the Windows
and Linux drivers have never supported Jumbos on this chipset and since
it is so old and rare it is not that big of a deal.
From FreeBSD
ok dlg@ awhile ago.
brad [Fri, 29 Aug 2008 22:59:56 +0000 (22:59 +0000)]
Take the PHY out of power down while attaching just before probing for
PHY. This helps with systems dual booting Windows XP/Vista where the
Windows drivers will put the PHY in power down mode when rebooting.
From FreeBSD
oga [Fri, 29 Aug 2008 15:14:59 +0000 (15:14 +0000)]
Build with mtrr support on amd64. This really should have gone in in
edmonton, but I forgot.
ok and extra testing phessler@.
oga [Fri, 29 Aug 2008 13:44:23 +0000 (13:44 +0000)]
Don't memcpy too far whem drm_realloc() is called with a smaller size.
This was never noticed since it's always used with a larger size.
Noticed by Stephane Marchesin.
oga [Fri, 29 Aug 2008 13:07:13 +0000 (13:07 +0000)]
Fix nvi's cscope support in the case that someone provided a filename
without a directory (e.g. :cscope add cscope.out). Found and fixed by
Paul Irofti, with help from me; Thanks!
upstream nvi also has this fix.
ok otto@.
damien [Fri, 29 Aug 2008 12:14:53 +0000 (12:14 +0000)]
move code to support Frequency-Hopping spread spectrum (FHSS) PHYs
to the Attic. nothing uses it in the tree and it is very unlikely
that something will use it one day.
the only driver supporting FHSS PHYs in the tree is ray(4) and it
does not use net80211.
jakob [Fri, 29 Aug 2008 11:47:49 +0000 (11:47 +0000)]
remove BIND's delegation-only from default config files.
ok djm@ millert@
reyk [Fri, 29 Aug 2008 11:15:32 +0000 (11:15 +0000)]
add support for software crypto to allow WPA/WPA2/RSN. it is disabled
for now because it needs more testing, but basic WPA/WPA2 and WEP
seems to work. to enable it, set the compiled-in ath_softcrypto
variable to 1.
this is based on a previous diff from damien@ with some changes to
disable the hardware crypto engine if softcrypto is enabled and to
keeps the hardware crypto code in place to allow later work on
hardware WPA/WPA2.
damien [Fri, 29 Aug 2008 10:52:41 +0000 (10:52 +0000)]
acx(4) supports Host AP mode too and does not handle PS STAs properly
(since it uses net80211 and does not manage dynamic parts of beacons).
spotted by jmc@
reyk [Fri, 29 Aug 2008 10:05:00 +0000 (10:05 +0000)]
add a HAL function to force disabling the hardware crypto engine
tobias [Fri, 29 Aug 2008 09:54:22 +0000 (09:54 +0000)]
Pre-commit checks (CVSROOT/commitinfo) are run before log message is
requested. Do this for GNU cvs compatibility, too.
ok joris
tobias [Fri, 29 Aug 2008 09:51:21 +0000 (09:51 +0000)]
Support for variable expansion in trigger files added.
ok joris
tobias [Fri, 29 Aug 2008 09:46:10 +0000 (09:46 +0000)]
If parsing of a command in one of the trigger files fails, treat this as an
error instead of ignoring it.
ok joris
damien [Fri, 29 Aug 2008 09:44:16 +0000 (09:44 +0000)]
new sentence, new line.
damien [Fri, 29 Aug 2008 09:30:23 +0000 (09:30 +0000)]
the atw(4) man page says:
"The driver lets you change to Host AP mode, but it does not work
and it probably never will."
so just remove the HOSTAP capability bit in the code and remove this
sentence.
damien [Fri, 29 Aug 2008 09:03:48 +0000 (09:03 +0000)]
add the following paragraph from the wi(4) manual page to the
CAVEATS section of all drivers claiming HostAP support:
"Host AP mode doesn't support power saving. Clients attempting to
use power saving mode may experience significant packet loss
(disabling power saving on the client will fix this)."
unfortunately (or not?), more and more clients are using PS mode by
default and will not work properly with an AP running OpenBSD.
proper handling of STAs in PS mode will come in net80211 but is is
very unlikely that all drivers will be ported.
otto [Fri, 29 Aug 2008 08:57:28 +0000 (08:57 +0000)]
fix comment: we use a pool for namei, so malloc considerations do not
apply; ok jsg@
damien [Fri, 29 Aug 2008 08:40:53 +0000 (08:40 +0000)]
garbage collect unused capabilities and configurations flags.
add two capabilities flags: IEEE80211_C_HT for HT STAs (802.11n)
and IEEE80211_C_APPMGT which indicates the capability for an AP
to buffer unicast and multicast traffic for STAs in PS mode.
all drivers claiming HostAP support should support that but the
truth is that none of them do.
most of them are still at the 802.11b-only era and do not update
dynamic parts of beacons or process frames from ic_pwrsaveq.
ajacoutot [Fri, 29 Aug 2008 08:17:28 +0000 (08:17 +0000)]
- prevent sysmerge from installing a virgin passwd and group
files (only allow delete or merge+install); requested by nick@
no objections from laurent@ krw@ oga@
ok sthen@ nick@
jmc [Fri, 29 Aug 2008 08:13:08 +0000 (08:13 +0000)]
MSIZE and MCLBYTES are now defined in sys/param.h; from Jordan Gordeev
ok thib
cnst [Fri, 29 Aug 2008 03:38:31 +0000 (03:38 +0000)]
support AMD Family 11h (Turion X2 Ultra et al); discussed with jsg@ and jmc@
oga [Thu, 28 Aug 2008 18:29:08 +0000 (18:29 +0000)]
Ignore xserver provided mmio address. We already have it. From git.
damien [Thu, 28 Aug 2008 17:56:24 +0000 (17:56 +0000)]
unbreak IEEE80211_DEBUG
brad [Thu, 28 Aug 2008 17:51:09 +0000 (17:51 +0000)]
Rewrite the multicast/promiscuous handling and in the process merge into
one function. Also fix reception of multicast packets when in promisuous
mode. The Realtek chips are a little different than most others in that
they will not receive multicast packets by enabling promiscuous mode
alone.
Tested by naddy@ and sthen@
ok naddy@
damien [Thu, 28 Aug 2008 16:02:14 +0000 (16:02 +0000)]
indent IF_PURGE. pointed out by brad@
no binary changes.
damien [Thu, 28 Aug 2008 16:00:21 +0000 (16:00 +0000)]
#undef IPW_DEBUG
fix a comment while i'm here.
pointed out by brad@
damien [Thu, 28 Aug 2008 15:55:37 +0000 (15:55 +0000)]
80 cols.
no binary changes.
damien [Thu, 28 Aug 2008 15:52:20 +0000 (15:52 +0000)]
WPA support for iwi(4).
some initial WMM bits too.
use license.template while i'm here.
brad [Thu, 28 Aug 2008 15:19:03 +0000 (15:19 +0000)]
Do not try to attach to rev 2 BCM4311/BCM4312 chipsets as they
require v4 firmware and the driver currently uses v3 firmware.
ok mglocker@
damien [Thu, 28 Aug 2008 15:08:38 +0000 (15:08 +0000)]
i've lost the IF_PURGE() bits in the process...
use license.template while i'm here.
damien [Thu, 28 Aug 2008 14:40:44 +0000 (14:40 +0000)]
WPA support for ipw(4).
Did a lot of cleanup while I was there.
brad [Thu, 28 Aug 2008 13:19:38 +0000 (13:19 +0000)]
some style fixes.
brad [Thu, 28 Aug 2008 13:10:54 +0000 (13:10 +0000)]
Fix a typo in the naming of the PPPoE input queues, ppoe -> pppoe.
ok canacar@
jmc [Thu, 28 Aug 2008 13:09:02 +0000 (13:09 +0000)]
some auacer(4) bits;
reyk [Thu, 28 Aug 2008 11:10:25 +0000 (11:10 +0000)]
encode the speed and link state in the local port key.
ok brad@
mikeb [Thu, 28 Aug 2008 10:29:35 +0000 (10:29 +0000)]
man page for the auacer(4) driver.
mikeb [Thu, 28 Aug 2008 10:21:22 +0000 (10:21 +0000)]
Driver for Acer Labs M5455 integrated sound found on recent Ultra workstations.
Suggestions and corrections by kettenis, ratchov, jakemsr. Thanks a lot!
Recording doesn't work at the moment.
ok kettenis ratchov
jmc [Thu, 28 Aug 2008 08:39:44 +0000 (08:39 +0000)]
typo; from hyjial
ok otto
jasper [Thu, 28 Aug 2008 05:18:26 +0000 (05:18 +0000)]
- enable uvideo on sgi, it's working for me with a quickcam e 3500 plus
ok mglocker@
oga [Thu, 28 Aug 2008 02:26:16 +0000 (02:26 +0000)]
Ignore userland provided batchbuffer start parameter, we can determine
it ourselves.
From drm git.
oga [Thu, 28 Aug 2008 01:51:52 +0000 (01:51 +0000)]
set sarea_priv to NULL to prevent possible access on next open/close.
From Robert Noland via drm git.
oga [Thu, 28 Aug 2008 01:34:18 +0000 (01:34 +0000)]
Kill the ifdefed out ttm interface stuff. The intel driver has
definitively gone for GEM so this will not be needed.
No binary change.
brad [Thu, 28 Aug 2008 01:17:29 +0000 (01:17 +0000)]
Add some more E1000 PCI ids.
From FreeBSD
oga [Thu, 28 Aug 2008 00:44:31 +0000 (00:44 +0000)]
This file was meant to be removed with the last commit, it's now empty
and unused.
oga [Thu, 28 Aug 2008 00:19:27 +0000 (00:19 +0000)]
replace usage of drm_memrange with extent(9). No functional change, but
should shrink the kernel somewhat. For some strange reason I was unaware
of this api when I pulled in these changes.
tested by myself and Paul de Weerd, thanks!
jmc [Wed, 27 Aug 2008 20:39:32 +0000 (20:39 +0000)]
document \user; from Kristaps D?onsons, originally from Mengzhu Wang
document comma separated entries; from millert
reorganise the page a bit to read better
ok millert
brad [Wed, 27 Aug 2008 20:38:59 +0000 (20:38 +0000)]
Remove a redundant define for RL_CFG2 and remove an incorrect comment.
jmc [Wed, 27 Aug 2008 20:20:13 +0000 (20:20 +0000)]
correct man sections; from Tobias Sarnowski
ok millert
jasper [Wed, 27 Aug 2008 20:12:00 +0000 (20:12 +0000)]
add an Xr to usb, now that we have usb on sgi
deraadt [Wed, 27 Aug 2008 17:33:32 +0000 (17:33 +0000)]
sync
deraadt [Wed, 27 Aug 2008 17:33:30 +0000 (17:33 +0000)]
another ups
mglocker [Wed, 27 Aug 2008 17:31:48 +0000 (17:31 +0000)]
Some __packed cleanup; Remove another obsolete pack from a none-hardware
struct and pack all other hardware structs which have been forgotten.
damien [Wed, 27 Aug 2008 10:58:01 +0000 (10:58 +0000)]
move definitions of channels attributes away from ieee80211.h such
that ieee80211.h contains only definitions that are part of the
802.11 standard and not constants used internally by net80211.
because channels attributes are exported to userland through the
radiotap BPF interface, add the definitions to ieee80211_radiotap.h
too (which must be kept in sync with what is used in net80211).
also, do not export combinations of channel attributes to userland
so that noone get stupid ideas.
damien [Wed, 27 Aug 2008 10:34:24 +0000 (10:34 +0000)]
pass FCS to radiotap.
damien [Wed, 27 Aug 2008 10:29:53 +0000 (10:29 +0000)]
fix IEEE80211_RADIOTAP_F_SHORTPRE
damien [Wed, 27 Aug 2008 10:03:47 +0000 (10:03 +0000)]
another IEEE80211_RADIOTAP_F_FCS candidate.
damien [Wed, 27 Aug 2008 10:01:18 +0000 (10:01 +0000)]
add IEEE80211_RADIOTAP_F_FCS radiotap flag from NetBSD.
use it in drivers that leave the 802.11 FCS in frames
passed to radiotap BPF. otherwise, userland has no way
to know if FCS is included or not as it depends on drivers.
this is required by some ports (aircrack).
requested by dhill@
damien [Wed, 27 Aug 2008 09:49:32 +0000 (09:49 +0000)]
hardware may pass us ACK or CTS frames in some cases, so use
IEEE80211_ACK_LEN instead of IEEE80211_MIN_LEN for ZYD_MIN_RXBUFSZ
and ZYD_MIN_FRAGSZ.
silence some warnings while i'm there.
change ZYD_FILTER_BSS to use the same value as the vendor's driver
that contains some magic (undocumented) bits.
damien [Wed, 27 Aug 2008 09:28:38 +0000 (09:28 +0000)]
the firmware is responsible for sending management frames, but
since we pass received management frames to net80211, net80211
may send replies (like deauth/disassoc), so we just call
IF_PURGE(&ic->ic_mgtq) in {ipw,iwi}_start just to be on the
safe side of things (so we don't leak mbufs).
damien [Wed, 27 Aug 2008 09:14:36 +0000 (09:14 +0000)]
override net80211's ic_node_alloc function to allocate a full
bwi_node structure (containing the rate control state).
because bwi(4) does not support HostAP or IBSS modes there is
no need to maintain a per-node rate control state, so we could
as well store it in bwi_softc but that will allow for future
improvements.
pointed out by Taylor R Campbell (campbell AT mumble DOT net)
on tech@
damien [Wed, 27 Aug 2008 09:05:03 +0000 (09:05 +0000)]
introduce new IEEE80211_STA_ONLY kernel option that can be set to
remove IBSS and HostAP support from net80211 and 802.11 drivers.
it can be used to shrink RAMDISK kernels for instance (like what
was done for wi(4)).
it also has the benefit of highlighting what is specific to IBSS
and HostAP modes in the code.
the cost is that we now have two code paths to maintain.
brad [Wed, 27 Aug 2008 08:41:46 +0000 (08:41 +0000)]
Fix length check when parsing PPPoE discovery phase packets.
Malicious PPPoE discovery packets could cause the kernel to
crash.
From canacar@ and inspired by the original fix from NetBSD.
ok canacar@
kjell [Wed, 27 Aug 2008 04:11:52 +0000 (04:11 +0000)]
add a C-c target, so we can use it as a prefix for keybindings.
Now you can do something like:
global-set-key "\^c\^c" compile
in your ~/.mg
martynas [Wed, 27 Aug 2008 02:08:34 +0000 (02:08 +0000)]
exposed too much, only expose prototype outside ifdef
deraadt [Wed, 27 Aug 2008 01:37:16 +0000 (01:37 +0000)]
delete unused pragma ident junk
martynas [Wed, 27 Aug 2008 00:40:38 +0000 (00:40 +0000)]
comment fixes
martynas [Wed, 27 Aug 2008 00:31:01 +0000 (00:31 +0000)]
fix 128-bit division. gcc mangled arguments when passing to the
__udivti3, because MUST_PASS_IN_STACK always returned 1 on amd64;
pr#5780
reported by Simon Kuhnle
tested by Simon Kuhnle, sthen@, brad@
double-checked & tweak from miod@
ok sthen@, brad@
jsg [Tue, 26 Aug 2008 22:49:09 +0000 (22:49 +0000)]
Fix an if (); typo. ok beck@
deraadt [Tue, 26 Aug 2008 22:25:20 +0000 (22:25 +0000)]
another ups to knock out of being a uhid
deraadt [Tue, 26 Aug 2008 22:25:08 +0000 (22:25 +0000)]
sync
deraadt [Tue, 26 Aug 2008 22:24:30 +0000 (22:24 +0000)]
new devs
mglocker [Tue, 26 Aug 2008 21:52:48 +0000 (21:52 +0000)]
Fix alignment panic on sparc64 by removing a wrong __packed. While
there reorder fields in the struct to make it shorter on 64 bit archs.
panic reported by jasper@
thanks to miod@ for helping me debug this down
kettenis [Tue, 26 Aug 2008 21:06:29 +0000 (21:06 +0000)]
Fix a couple of problems that may make gem(4) get stuck:
1. If bus_dmamap_load_mbuf() fails because there are not enough
segments in the map, defrag the mbuf.
2. If there are not enough free (hardware ring) descriptors, set
IFF_OACTIVE and keep the packet on the queue.
3. If there is some other resource starvation that makes
bus_dmamap_load_mbuf() or defragmentation fail, drop the packet.
Don't set IFF_OACTIVE, since the Tx ring could be empty and we'd be
stuck.
4. Only pass packets that are actually handed off to the hardware to
BPF. Do so before handing them off to the hardware to make sure
the packet isn't freed behind our back.
ok dlg@
miod [Tue, 26 Aug 2008 19:46:23 +0000 (19:46 +0000)]
Stylus pressure sensor is active low, so invert the logic converting it
to button #5.
kettenis [Tue, 26 Aug 2008 19:43:05 +0000 (19:43 +0000)]
Better strategy to single out onboard bge(4)'s on sparc64 machines.
ok deraadt@
miod [Tue, 26 Aug 2008 18:36:21 +0000 (18:36 +0000)]
When using the not-documented-but-ought-to 'boot sd(...)file' syntax (e.g.
when mopbooting the bootblocks because one trashed the ones on disk by
accident), be sure to use the proper BDEV_SDx rpb device type value,
depending on the type of onboard controller. Crank version.
martynas [Tue, 26 Aug 2008 18:29:12 +0000 (18:29 +0000)]
- add F
- make inf INF nan NAN comply to standards (eEfFgG)
- extend man page bits
ok millert@. w/ a man page tweak and ok jmc@
martynas [Tue, 26 Aug 2008 17:56:30 +0000 (17:56 +0000)]
don't zero-pad special values. ok millert@
kettenis [Tue, 26 Aug 2008 15:39:27 +0000 (15:39 +0000)]
Plug potential memory leak.
"looks sane to me" otto@, ok miod@
henning [Tue, 26 Aug 2008 12:19:01 +0000 (12:19 +0000)]
we need to call pf_pkt_addr_changed here too. found by david
henning [Tue, 26 Aug 2008 12:18:01 +0000 (12:18 +0000)]
call pf_pkt_addr_changed instead of manually clearing the pf state key ptr
henning [Tue, 26 Aug 2008 12:17:10 +0000 (12:17 +0000)]
introduce a function to be called when addressing information has changed,
pf_pkt_addr_changed. atm just clears the state key pointer.
calling this is cleaner than having other parts of the stack clearing
pointers in the pf part of the mbuf packet header directly.
jmc [Tue, 26 Aug 2008 08:59:49 +0000 (08:59 +0000)]
libbfd and libopcodes no longer built;
brad [Tue, 26 Aug 2008 07:43:03 +0000 (07:43 +0000)]
Update comment to mention the 8211C.
brad [Tue, 26 Aug 2008 07:40:32 +0000 (07:40 +0000)]
Mention the 8211C.